Final Thoughts

Overview

Teaching: 5 min
Exercises: 10 min

Questions

Objectives

Is machine learning really the answer?

We started this lesson with a toy problem: you were hosting a party for your friends, and had a giant bowl of mixed candies. You realized that one of your friends is allergic to peanuts, and that they might mistakenly take a peanut M&M’s from the bowl of candies. You want to keep your friend safe, so you decided to separate out the non-peanut candies from the peanut ones using a machine learning algorithm (specifically, a supervised classification one).

As we learned how to use machine learning to solve this problem, some ethical questions came up. However, there is an important one that we missed and that should always be asked: Should we even be using machine learning to separate the candy in the first place? Why or Why not?

Peanut allergies can be very serious, and some people will have an allergic reaction to food that has simply come into contact with peanuts. Thus, even if our machine learning algorithm was perfect, it might not be the solution to the problem. Instead, the safest thing to do would be to buy a completely new bag of candy that has the guarantee “Peanut Free” symbol, and give this bag to your friend when they arrive to the party.

Key Points